Does Phrasly bypass Turnitin? It is the one company here that gave you a number to check — and checking it is what this page is for.
The only tool that bypassed Pangram and Turnitin in 2026.
Most humanizers clear one detector and get caught by the other. StealthWriter is the one that gets past both — run Ghost 5.2 Pro at level 7–8, section by section, and re-check before you submit.
Try StealthWriter →Key Takeaways
- Phrasly is the only vendor in this market that has published a dated, numbered Turnitin result: “100 documents…all 100 returned a 0% AI score.”
- It is filed in a versioned engineering changelog, timed to Turnitin’s own product update, and hyperlinked to Turnitin’s guides. It has the shape of an audit.
- It publishes no screenshot of any of the 100 reports, and its own preamble downgrades the page to “our internal evaluations.”
- 0% survives the usual falsification test — Turnitin does display zero. It fails a different one: 100 instructor-side reports in one month, through an account it never names.
- Its Terms say, in capitals: “WE CANNOT GUARANTEE THAT THE TEXT GENERATED BY OUR SERVICES WILL ALWAYS PASS SUCH DETECTORS.”
- Section 31: it “does not support using our technology to circumvent educational AI detection systems” — four clicks from a page boasting it retrains whenever Turnitin updates.
- Refunds require “absolutely no usage…whatsoever” — zero documents processed across the account’s entire history.
The only tool that bypassed Pangram and Turnitin in 2026.
Most humanizers clear one detector and get caught by the other. StealthWriter is the one that gets past both — run Ghost 5.2 Pro at level 7–8, section by section, and re-check before you submit.
Try StealthWriter →Does Phrasly bypass Turnitin?
Phrasly says yes, and unlike every other vendor in this cluster it says so with a figure, a date and a sample size. That deserves to be taken seriously enough to examine properly.
The entry sits in a versioned changelog, dated 27 August 2025, headed “Turnitin product update: compatibility review”. It is timed to Turnitin’s own release, and it hyperlinks to Turnitin’s product-update notes.
Dated entry. Stated sample size. Pass-fail result. Upstream citation. It has the shape of an engineering audit, and that shape is doing most of the persuasive work.
This page reports no first-party test; everything below was fetched on 31 August 2026.
First, the test it passes
There is a quick falsification test that catches most fake Turnitin claims, and Phrasly’s survives it.
Turnitin’s documentation says “no score or highlights are attributed for AI detection scores in the 1% to 19% range” — below 20% the indicator shows an asterisk and no number, because false positives are more common in that band. So a vendor claiming a Turnitin AI score of 3% or 12% is describing something Turnitin does not render.
Phrasly claims 0%. Zero is a value Turnitin displays. The claim is display-legal, and we say so plainly because most of its competitors’ numbers are not.
Now the test it does not pass
| Who | Can see a Turnitin AI writing score? |
|---|---|
| Instructors at licensed institutions | Yes |
| Students, even at their own institution | No — the report is not visible to them |
| Anyone buying an individual licence | No such licence is sold |
| Draft Coach users | No — it has no AI detection |
| A humanizer vendor | No lawful route |
Turnitin’s help centre states it “does not sell individual licenses or single-use subscriptions”, and that “AI writing detection reports are not visible to students”.
Read the claim against that. One hundred documents checked against Turnitin, inside one month. Each check produces an AI writing report that only an instructor at a licensed institution can open. To generate the line, someone would have needed a hundred instructor-side reports and a route to them.
That is the finding. Phrasly borrowed the form of an audit — a dated entry, a sample size, a pass rate, a link to the upstream vendor’s release notes — for a result that is structurally unobtainable through any route it is entitled to use. The form is what makes it persuasive. The substance is one sentence.
Its own Terms say the opposite
Section 21 of the Terms of Service, last updated 10 April 2026 — seven months after the changelog entry — reads, in the original capitals:
“DUE TO THE CONSTANTLY EVOLVING NATURE OF AI DETECTION TOOLS AND THEIR POTENTIAL FOR FALSE POSITIVES, WE CANNOT GUARANTEE THAT THE TEXT GENERATED BY OUR SERVICES WILL ALWAYS PASS SUCH DETECTORS… EVEN WITH OUR BEST EFFORTS TO ENSURE HUMAN-LIKE OUTPUT, WE CANNOT BE HELD RESPONSIBLE IF OUR AI-GENERATED TEXT IS FLAGGED BY AI DETECTORS.”
And section 31: “Phrasly.ai does not support using our technology to circumvent educational AI detection systems when they are used to prevent academic misconduct.”
Four clicks away, its technology page says: “Every time Turnitin, GPTZero, or any other checker updates, our dedicated team is already on it. They monitor changes daily and retrain our models to stay ahead.” Turnitin is an educational AI detection system used to prevent academic misconduct. Those two sentences are on the same domain.
What does independent work say?
Pangram’s DAMAGE paper audits 19 humanizers and rates Phrasly “L3” — the lowest of three quality tiers — describing its output verbatim as “Poor quality sentences. Much worse in some texts rather than others.” Phrasly is absent from Pangram’s August 2025 benchmark‘s 20 scored tools.
Two caveats must travel with that or the citation is unfair. The paper states it classifies tools “based on faithfulness and fluency, not on their effectiveness at bypassing AI detectors”, and it separately observes that less fluent output can be harder to detect. An L3 rating is information about the writing you would be submitting, not a verdict on evasion. And neither Pangram document tests any tool against Turnitin.
There are also no Turnitin screenshots on phrasly.ai. Its image assets are product graphics, three how-it-works steps, stock testimonial faces, and university logos including Berkeley, Cambridge, MIT and Oxford — on a site whose ethics page is headed “We stand with educators”.
What can be checked on Turnitin’s side?
Nothing that would confirm or refute the 100-document claim. Turnitin shipped paraphrasing detection in July 2024 and dedicated bypasser detection in August 2025 — the very update Phrasly’s changelog entry responds to — and has published no accuracy figure for either since. On 4 August 2026 it merged the AI-paraphrased highlight into AI-generated, so even instructors can no longer see how much of a flag came from bypasser detection.
Independently, Hadra, Cambridge and Mesbah (February 2026, 192 texts) measured Turnitin at 0.61 overall accuracy and 0.51 recall, performing poorly on hybrid human-and-AI text. Weber-Wulff et al. (2023) put field-wide accuracy on machine-paraphrased text at 26%, with Turnitin the best of 14 tools tested.
What should you take from this?
Phrasly gave this market its only falsifiable Turnitin number, and that is genuinely more than anyone else offered. Held up to the light, the number requires a hundred instructor-side reports obtained through an unnamed route, is unaccompanied by a single artefact, is described by its own publisher as an internal evaluation, and is contradicted by the company’s own Terms seven months later.
It is also worth being clear about what would change our view: a screenshot with a submission ID, or a named institution. Neither exists.
Turnitin’s own guidance says its model “may misidentify human-written, AI-generated, and AI-paraphrased text, so it should not be used as the sole basis for adverse actions against a student”, and four universities have disabled it — Vanderbilt, Waterloo, Curtin and Washington State. And Phrasly’s refund policy requires “absolutely no usage…whatsoever”, so finding out costs you the subscription.
If your real worry is being wrongly flagged on your own writing, keep your drafting history — it costs nothing and it survives a detector being wrong. It matters most for the writers detectors treat worst: ESL writers and AI detection and false positives for neurodivergent students. Check before you submit with the best pre-submission check and the Turnitin self-check, and strip the obvious tells by hand first via what to remove before using an AI humanizer.
For the same question against Pangram, see does Phrasly bypass Pangram. Other Turnitin cells: Clarity Bubble, WriteHuman, Ryne AI, UnAIMyText, AuraWrite, Text Polish, HueWrite, AiTextools, RewriteIQ, GenZWrite, GPT Scrambler, Rephrasy and StealthWriter.
Get the free prompt pack instead
The manual rewriting prompts that lower AI signals without a subscription — and without betting a submission on a number nobody has reproduced. Free, no card.
Frequently asked questions
Does Phrasly bypass Turnitin?
Phrasly says yes and gives a number, which no other vendor in this market does. Its changelog reports 100 documents checked against Turnitin in August 2025, all returning 0% AI. That claim is dated, specific and falsifiable in principle — and it rests entirely on one sentence of self-attested text, with no screenshot of any of the 100 reports and no account named.
What is the exact claim?
From a changelog entry dated 27 August 2025, headed “Turnitin product update: compatibility review”: “Following Turnitin’s latest product update, we re-evaluated the current humanizer model. No model changes were required.” Then: “Regression test: 100 documents processed through the humanizer were checked against Turnitin in August 2025; all 100 returned a 0% AI score.” The entry links to Turnitin’s own product-update notes.
Does 0% fail the fake-screenshot test?
No, and that is worth stating fairly. Turnitin suppresses any AI figure between 1% and 19%, showing an asterisk instead, so a claimed 3% or 12% would be self-refuting. Zero is a value Turnitin does render. The claim is display-legal. It fails on a different axis entirely.
What is the problem, then?
Access. Turnitin states it “does not sell individual licenses or single-use subscriptions”, its AI writing report is instructor-facing and “not visible to students”, and Draft Coach has no AI detection. To produce that line, somebody would have needed 100 separate instructor-side Turnitin AI reports inside one month. Phrasly names no institution, no account and no reviewer, and publishes no screenshot of a single one.
Do its own Terms support the claim?
They contradict it. Section 21, in the original capitals: “DUE TO THE CONSTANTLY EVOLVING NATURE OF AI DETECTION TOOLS AND THEIR POTENTIAL FOR FALSE POSITIVES, WE CANNOT GUARANTEE THAT THE TEXT GENERATED BY OUR SERVICES WILL ALWAYS PASS SUCH DETECTORS.” The Terms were last updated 10 April 2026, after the changelog entry.
What about its ethics position?
Section 31 states that Phrasly “does not support using our technology to circumvent educational AI detection systems when they are used to prevent academic misconduct.” Its technology page says: “Every time Turnitin, GPTZero, or any other checker updates, our dedicated team is already on it. They monitor changes daily and retrain our models to stay ahead.” Turnitin is an educational AI detection system.
Is Phrasly in independent research?
Yes, and not favourably. Pangram’s DAMAGE paper rates Phrasly “L3” — the lowest of three quality tiers — describing its output as “poor quality sentences. Much worse in some texts rather than others.” Two caveats: the paper rates faithfulness and fluency explicitly “not… effectiveness at bypassing AI detectors”, and it does not test any tool against Turnitin.
What does it cost, and can I get a refund?
Three-day access is $2.00; Unlimited is $10.99 a month billed annually. The refund policy is the strictest in this cluster: 14 days, but “only available if there has been absolutely no usage on your account whatsoever… Any usage of our services, at any point, immediately voids refund eligibility.” The $2 trial fee is non-refundable, and a chargeback voids both your access and your refund eligibility.
